This table contains 7,927 entries, which is the current published size of the ISO 639-3 registry. That's dramatically larger than the other two parts of the standard: ISO 639-1 covers roughly 183 major languages with two-letter codes, and ISO 639-2 covers around 487 languages with three-letter bibliographic and terminology codes. ISO 639-3 covers both of those plus almost everything else: small living languages that never made it into the smaller lists, extinct languages documented only in historical texts, historical-stage languages, constructed languages, and a handful of special catch-all codes. Of the 7,927 entries here, 7,082 are living languages, 602 are extinct, 215 are historical-stage languages, and 24 are constructed. No other part of the ISO 639 standard attempts this scope, which is why SIL maintains 639-3 as a standalone, continuously updated registry rather than a fixed list.

A macrolanguage is a single ISO 639-3 code that stands in for a group of closely related, mutually distinguishable individual languages that speakers or governments commonly refer to by one umbrella name. Arabic and Chinese are the two clearest examples: many speakers and most everyday contexts treat "Arabic" or "Chinese" as one language, but linguistically they're clusters of varieties that are separately catalogued as individual ISO 639-3 languages. This registry contains 63 macrolanguage codes in total. Cross-referencing SIL's official macrolanguage-mapping table, the Arabic macrolanguage, coded ara, maps to 30 individual language codes covering varieties like Egyptian Arabic, Gulf Arabic, and Moroccan Arabic, and the Chinese macrolanguage, coded zho, maps to 19 individual codes covering varieties like Mandarin, Cantonese, and Hakka. Requesting "Arabic" or "Chinese" without further detail is, in ISO 639-3 terms, requesting the macrolanguage code rather than a specific one of its 30 or 19 individual varieties.

602 of the 7,927 entries in this table are classified as Extinct in SIL's Language Type field: languages with no remaining native speakers, documented in historical or field-linguistics records but no longer in active use anywhere. That works out to roughly 7.6% of the registry, a small share next to the 7,082 living languages, but still a far larger and more deliberate accounting of extinct languages than either ISO 639-1 or ISO 639-2 attempts. Cataloging languages that have died out is one of the main reasons ISO 639-3 exists as a separate, more exhaustive part of the standard.

SIL International is the registration authority for ISO 639-3, a role formalized when the standard was published in 2007. Anyone can submit a change request through SIL's public change-request process to add a new code, split or merge existing ones, retire a code, or correct a name. Requests are reviewed against documented linguistic evidence, opened for public comment, and then approved or declined; approved changes are folded into the next annual release of the code tables, which is the same release cycle this page's dataset comes from.

Not for hreflang. Google's own documentation for hreflang is explicit: only language codes listed in ISO 639-1 are supported, with region codes from ISO 3166-1, and no fallback to ISO 639-2 or ISO 639-3. That means most of the languages listed in this table, the ones that never received an ISO 639-1 code, can't be targeted with an hreflang tag at all, even though each has a valid, published ISO 639-3 identifier. Whether other software or platforms accept ISO 639-3 codes inside a BCP 47 tag varies by implementation; the code itself is valid, but that doesn't guarantee any particular piece of software will recognize it.

No. ISO 639-3 identifies a language and nothing else, the same as ISO 639-1 and ISO 639-2. ISO 3166 is a separate standard that identifies countries and regions, such as US for the United States or FR for France, not languages. The two standards are often combined, not confused, in real systems: a locale tag like en-US pairs an ISO 639-1 language subtag with an ISO 3166 region subtag to say "English as used in the United States." An ISO 639-3 code can, in principle, be combined with an ISO 3166 code the same way under BCP 47, but as the previous answer notes, Google's hreflang doesn't recognize that combination; it requires an ISO 639-1 language code specifically.

No. Boostlingo's list of supported languages is organized by language name, not by ISO code, and it lists some individual dialects separately where Boostlingo tracks that level of detail.